Aiserver Manufacturer
1 / 5
| Introduction | The Atlas 300I Duo 96GB is a dual-chip AI inference acceleration card, designed specifically for large language models (LLM) and video analysis. It offers 96GB memory and 280 TOPS of INT8 computing power. |
| Benchmark Product | Comparable to T4 |
| Computing Power | INT8: 280 TOPS | FP16: 140 TFLOPS |
| VRAM Type | HBM2E (High Bandwidth Memory, approx. 30% higher bandwidth than HBM2) |
| Fabrication Process | TSMC 7nm FinFET with 2.5D CoWoS advanced packaging |
| DirectX Support | Compatible with DirectX 12 (including 12_1 features) |
| Memory Frequency | Effective 3600MHz, 800GB/s bandwidth |
| Bus Width | 2048-bit |
| Output Interface | PCIe Gen4.0 x16 (focused on data center inference) |
| Form Factor | FHFL dual-slot (1064g) |
Cyber Resilient Architecture ensures security at every phase of the lifecycle. Silicon-based root of trust anchors end-to-end boot resilience with Multi-Factor Authentication.
Increase efficiency with autonomous collaboration. Simplify and automate management through secure consoles and integrated intelligent remote access.
Designed for energy efficiency and reduced carbon footprint. Includes innovative options for recycling and professional legacy system retirement.
Designed for 24x7 enterprise data center operations, it features enterprise-grade components and energy-efficient hardware optimized for deployment at scale. Supports Secure Boot with Hardware Root of Trust technology, providing an extra layer of security that meets NEBS Level 3 standards.
The passively cooled, energy-efficient design fits a wide range of Certified systems from OEM vendors and is easy to integrate into existing data center infrastructure.
